Method and system for simulating broadband equivalent model

Abstract

The invention discloses a method and a system for simulating a broadband equivalent model. The method comprises the following steps: acquiring power grid parameters and a frequency range, calculating frequency domain characteristics of a core element, shrinking a network boundary through a Gaussian elimination method, and calculating a multi-port admittance matrix; fitting a multi-port admittance matrix, converting frequency domain characteristics into a time domain sequence through inverse fast Fourier transform, constructing a coefficient matrix, estimating a matrix rank based on singular value decomposition, and determining an optimal model order, namely a broadband equivalent model, by combining the number of real number poles; splitting the broadband equivalent model into a real number pole fraction sub-block, a constant term / first term sub-block and a plurality of pole fraction sub-blocks, performing passivity correction and verification on each fraction sub-block, and determining a corrected broadband equivalent model; and compressing the corrected broadband equivalent model, and splitting the broadband equivalent model into a plurality of sub-modules for parallel simulation.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of electromagnetic transient simulation of power system, and more particularly, to a method and system for simulating a wideband equivalent model. BACKGROUND

[0002] The safety and stability of power grid operation depend on accurate simulation analysis. Especially after a high proportion of new energy and power electronic devices are connected, the time scale of system dynamic response extends to microseconds, and electromagnetic transient simulation becomes the core means to characterize complex dynamic characteristics. However, electromagnetic transient simulation has a large amount of calculation, which is difficult to be directly applied to large-scale power grids, and usually requires equivalent network to represent non-concerned areas.

[0003] The conventional network equivalent method (such as Norton equivalent) is only based on the fundamental frequency characteristics, which will lose the high frequency characteristics of the original network, resulting in simulation distortion of key dynamic processes such as DC commutation failure and non-fundamental wave coupling in AC-DC hybrid systems. Although the existing wideband equivalent (FDNE) technology can describe frequency-dependent characteristics, it has three major bottlenecks: first, the wideband characteristics scanning efficiency of high-order multi-port power grid is low, which is difficult to adapt to the frequency characteristic sampling demand of large-scale power grid; second, the model order needs to be manually tried in the frequency fitting process, and the accuracy and efficiency are difficult to balance; third, the passive correction method either has low computational efficiency or cannot guarantee the passivity in the whole frequency domain, resulting in unstable simulation; in addition, the real-time simulation efficiency of the equivalent circuit is insufficient, which limits its application in large-scale power grids.

[0004] Therefore, it is of great significance to develop an efficient and accurate wideband characteristic analysis module to realize the integrated processing of wideband characteristic scanning, fitting, correction and simulation acceleration, so as to improve the electromagnetic transient simulation capability of power system. SUMMARY

[0005] According to the present application, a method and system for simulating a wideband equivalent model are provided to solve the technical problems of low efficiency of wideband characteristic analysis of high-order multi-port power grid, tedious identification of frequency fitting order, insufficient stability of passive correction, and low simulation efficiency.

[0006] According to a first aspect of the present application, a method for simulating a wideband equivalent model is provided, comprising:

[0007] obtaining power grid parameters and a frequency range, calculating the frequency domain characteristics of core elements, based on the frequency domain characteristics, contracting the network boundary by Gaussian elimination method, and calculating the multi-port admittance matrix;

[0008] fitting the multi-port admittance matrix, converting the frequency domain characteristics into time domain sequence by inverse fast Fourier transform, constructing a coefficient matrix according to the time domain sequence and estimating the matrix rank based on singular value decomposition, and determining the wideband equivalent model in combination with the number of real poles.

[0009] Splitting the wideband equivalent model into real pole fraction, constant / linear term part and complex pole fraction sub-block, correcting passivity of each fraction sub-block, determining the corrected wideband equivalent model;

[0010] Compressing the corrected wideband equivalent model and splitting it into multiple sub-modules for parallel simulation.

[0011] Optionally, the power grid parameters and the frequency range are obtained, the frequency domain characteristics of the core element are calculated, the network boundary is contracted through Gaussian elimination based on the frequency domain characteristics, and the multi-port admittance matrix is calculated, including:

[0012] The power grid parameters and the frequency range are obtained, the corresponding frequency domain characteristic model is established for different core elements, and the frequency domain characteristics of the core element are calculated;

[0013] The power grid topology and parameters are read, the positive / negative / zero sequence admittance matrix is constructed, the influence of the load and the DC device is considered, the port frequency domain characteristics are solved through LU decomposition, and the full frequency domain port admittance sampling data is output.

[0014] Optionally, the multi-port admittance matrix is fitted, the frequency domain characteristics are converted into time domain sequences through inverse fast Fourier transform, the coefficient matrix is constructed according to the time domain sequences and the matrix rank is estimated based on singular value decomposition, and the wideband equivalent model is determined in combination with the number of real poles, including:

[0015] Based on the relaxation auxiliary function and through QR decomposition to optimize the least square solution efficiency, the rational function of the multi-port admittance matrix Y(s) is fitted, and the expression is:

[0016]

[0017] Wherein, Ri is the residue matrix, a i is the pole, D and E are the constant term and linear term matrix;

[0018] The constant term, the linear term and the real pole contribution in the frequency domain characteristics are removed, the negative frequency conjugate symmetry data is completed, the time domain sequences are obtained through IFFT, the order is identified through singular value threshold, and the final order is the sum of the number of real poles and the rank.

[0019] Optionally, the wideband equivalent model is split into real pole fraction, constant / linear term part and complex pole fraction sub-block, each fraction sub-block is corrected for passivity, and the corrected wideband equivalent model is determined, including:

[0020] The complex pole fraction is corrected according to the following formula

[0021]

[0022] by restricting Ensure that the sub-block is passive, and integrate all sub-blocks to get the full frequency domain passive model;

[0023] By Hamilton matrix eigenvalue detection, ensure that the real part eigenvalue of the corrected wide frequency equivalent model is non-negative, and the out-of-bound frequency band is eliminated.

[0024] Optionally, the corrected wide frequency equivalent model is compressed and split into multiple sub-modules for parallel simulation, including:

[0025] Based on the singular value decomposition of the residue matrix, the rank (N is the number of ports) of the residue matrix is compressed, and R k = U∑V T Simplify the model structure and reduce the number of state variables;

[0026] The corrected wide frequency equivalent model is split into k sub-modules, each sub-module independently calculates the equivalent admittance and historical current, and the single time step calculation amount is reduced through parallel calculation, and the calculation amount of a single sub-module is

[0027] O component = 2nN 2 / k+N 2 + 2nN / k.

[0028] According to another aspect of the present application, a system for simulating a wide frequency equivalent model is also provided, which comprises:

[0029] A multi-port admittance matrix module is used to obtain power grid parameters and frequency range, calculate frequency domain characteristics of core elements, contract network boundaries based on the frequency domain characteristics through Gaussian elimination method, and calculate multi-port admittance matrix;

[0030] A wide frequency equivalent model module is used to fit the multi-port admittance matrix, convert the frequency domain characteristics into time domain sequence through inverse fast Fourier transform, construct a coefficient matrix according to the time domain sequence and estimate the rank of the matrix based on singular value decomposition, determine the wide frequency equivalent model in combination with the number of real poles;

[0031] A corrected wide frequency equivalent model module is used to split the wide frequency equivalent model into real pole fraction, constant term / linear term part and complex pole fraction sub-block, correct the passivity of each fraction sub-block, and determine the corrected wide frequency equivalent model;

[0032] A wide frequency equivalent model simulation module is used to compress the corrected wide frequency equivalent model and split it into multiple sub-modules for parallel simulation.

[0033] Optionally, the multi-port admittance matrix calculation module comprises:

[0034] The frequency domain characteristic calculation sub-module is configured to obtain power grid parameters and a frequency range, establish a corresponding frequency domain characteristic model for different core elements, and calculate the frequency domain characteristics of the core elements;

[0035] The output admittance sample data sub-module is configured to read power grid topology and parameters, construct a positive / negative / zero sequence admittance matrix, take into account the influence of loads and DC devices, solve port frequency domain characteristics through LU decomposition, and output full-frequency port admittance sample data.

[0036] Optionally, the wideband equivalent model determination module comprises:

[0037] The fitted admittance matrix sub-module is configured to fit a rational function of the multi-port admittance matrix Y(s) based on a relaxation auxiliary function and through QR decomposition to optimize the least square solution efficiency, and the expression is:

[0038]

[0039] wherein, Ri is a residue matrix, a i is a pole, and D and E are constant and first-order matrices;

[0040] The order identification sub-module is configured to remove the constant, first-order and real pole contributions in the frequency domain characteristics, complete the negative frequency conjugate symmetry data, obtain a time domain sequence through IFFT, identify the order through a singular value threshold, and the final order is the sum of the number of real poles and the rank.

[0041] Optionally, the wideband equivalent model correction module comprises:

[0042] The complex pole fraction correction sub-module is configured to correct the complex pole fraction according to the following formula

[0043]

[0044] By constraining to ensure that the sub-block is passive, and the full-frequency domain passive model is obtained by integrating all the sub-blocks.

[0045] The matrix eigenvalue detection sub-module is configured to detect the eigenvalues of the Hamilton matrix to ensure that the eigenvalues of the real part of the corrected wideband equivalent model are non-negative, and the out-of-bound frequency band is eliminated.

[0046] Optionally, the wideband equivalent model simulation module comprises:

[0047] The compressed residue matrix sub-module is configured to compress the residue matrix with a rank (N is the number of ports) based on singular value decomposition of the residue matrix, and obtain R k=U∑V T Simplify the model structure and reduce the number of state variables;

[0048] The corrected broadband equivalent model is divided into k sub-modules. Each sub-module independently calculates the equivalent admittance and historical current. Parallel computation is used to reduce the computational cost per time step, with the computational cost of a single sub-module being [missing information].

[0049] O component =2nN 2 / k+N 2 +2nN / k.

[0050] According to another aspect of the invention, a computer-readable storage medium is also provided, on which a computer program is stored, characterized in that, when executed by a processor, the program implements the steps of the method as described in any one of the claims.

[0051] According to another aspect of the present invention, an electronic device is also provided, comprising: the aforementioned computer-readable storage medium; and one or more processors for executing a program in the computer-readable storage medium.

[0052] This enables precise sampling of port frequency characteristics of large-scale multi-port power grids; solves the problem of manually specifying the order in vector fitting methods; ensures that wideband equivalent circuits are passive across the entire frequency domain and improves correction efficiency; develops compression and block-based parallel simulation techniques to improve the real-time simulation efficiency of wideband equivalent circuits; and constructs an integrated analysis module to achieve integrated operation of wideband characteristic scanning, fitting, correction, and simulation acceleration. Attached Figure Description

[0053] Exemplary embodiments of the present invention can be more fully understood by referring to the following figures:

[0054] Figure 1 This is a flowchart illustrating a method for simulating a broadband equivalent model as described in this embodiment;

[0055] Figure 2 This is a flowchart of the broadband characteristic scanning procedure described in this embodiment;

[0056] Figure 3 This is a schematic diagram of the voltage-type and current-type disturbance injection modules developed based on ADPSS as described in this embodiment;

[0057] Figure 4 This is a diagram of the harmonic injection element of the broadband scanning module described in this embodiment;

[0058] Figure 5 This is the impedance calculation parameter page for the wideband scanning module described in this embodiment;

[0059] Figure 6A schematic diagram of a system for simulating a broadband equivalent model according to an embodiment. DETAILED DESCRIPTION

[0060] Reference will now be made to the drawings to describe the exemplary embodiments of the present application in detail. The present application may, however, be embodied in many different forms and should not be construed as limited to the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and fully convey the scope of the application to those skilled in the art. Like reference numerals refer to like elements throughout the specification. It will be understood that when an element is referred to as being "on" another element, it can be directly on the element or intervening elements can also be present. In addition, it will also be understood that when an element is referred to as being "beneath" or "under" another element, it can be directly beneath the element or intervening elements can also be present. The terminology used herein is for the purpose of describing particular embodiments only and is not intended to be limiting of the present application. As used herein, the singular forms "a", "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be understood that the terms "comprises" and / or "comprising," when used in this specification, specify the presence of stated features, integers, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, and / or groups thereof.

[0061] Unless otherwise defined, all terms (including technical and scientific terms) used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. It will be further understood that terms, such as those defined in commonly used dictionaries, should be interpreted as having a meaning that is consistent with their meaning in the context of the relevant art and the present disclosure, and will not be interpreted in an idealized or overly formal sense unless expressly so defined herein.

[0062] According to a first aspect of the present application, there is provided a method 100 for simulating a broadband equivalent model, referring to Figure 1 The method 100 comprises:

[0063] S101: obtaining power grid parameters and frequency range, calculating frequency domain characteristics of core elements, based on the frequency domain characteristics, shrinking network boundary by Gaussian elimination method, calculating multi-port admittance matrix;

[0064] S102: fitting the multi-port admittance matrix, converting the frequency domain characteristics into time domain sequence by inverse fast Fourier transform, constructing coefficient matrix according to the time domain sequence and estimating matrix rank based on singular value decomposition, combining the number of real poles to determine the broadband equivalent model;

[0065] S103: splitting the broadband equivalent model into real pole fraction, constant term / linear term part and complex pole fraction sub-block, correcting each fraction sub-block for passivity, and determining the corrected broadband equivalent model;

[0066] S104: compressing the corrected broadband equivalent model and splitting it into multiple sub-modules for parallel simulation.

[0067] Specifically, the present application provides a method for simulating a broadband equivalent model, including broadband characteristic scanning, frequency fitting and order identification, passivity correction, simulation efficiency improvement and integrated interface, specifically as follows:

[0068] 1) broadband characteristic scanning, referring to Figure 2 and Figure 3 as shown:

[0069] a. Frequency range and sampling strategy, refer to Figure 4 and Figure 5 As shown: based on the electromagnetic transient simulation accuracy requirements, select 0-2500Hz as the core frequency range, use the logarithmic domain equal proportion incremental method to set 500 sampling points, and consider the low and high frequency characteristics capture.

[0070] b. Element frequency characteristic calculation: for the core elements such as generators, lines, and transformers, a simplified frequency characteristic model is established. For example, the sub-transient impedance frequency-dependent model is used for the generator, and the expression is

[0071] y(f) = 1 / (R g +jX g ·f / f0)

[0072] Where f0 is the base frequency, R g , X g are the base frequency parameters; the line uses the π-type frequency-dependent model, which considers the frequency dependence of resistance, inductance, and capacitance.

[0073] c. Port characteristic scanning: based on the ADPSS platform (full digital simulation platform of power system), the network boundary is contracted by the Gaussian elimination method, and the multi-port admittance matrix Y(s) is calculated. The process includes: reading the grid topology and parameters, constructing the positive / negative / zero sequence admittance matrix, considering the influence of load and DC devices, solving the port frequency characteristics by LU decomposition, and outputting the full frequency domain port admittance sampling data.

[0074] 2) Frequency fitting and order identification

[0075] a. Improved vector fitting method: use relaxation auxiliary function to speed up the convergence of poles, and through QR decomposition to optimize the least squares solving efficiency, realize the rational function fitting of multi-port admittance matrix Y(s), and the expression is

[0076]

[0077] Where R i is the residue matrix, a i is the pole, D and E are constant and first-order matrices.

[0078] b. Order identification based on Prony analysis: The frequency domain characteristics are converted into time domain sequences by inverse fast Fourier transform (IFFT), the coefficient matrix is constructed and the matrix rank is estimated based on singular value decomposition, and the optimal model order is determined combined with the number of real poles. The specific steps are as follows: remove the constant term, the first term and the contribution of real poles in the frequency domain characteristics, complete the negative frequency conjugate symmetry data, obtain the time domain sequence through IFFT, identify the order through the singular value threshold (0.9999-0.99999), and the final order is the sum of the number of real poles and the rank.

[0079] 3) Passive correction

[0080] a. Block SDP correction strategy: The wideband equivalent model is divided into real pole fraction, constant term / first term part and complex pole fraction sub-block, and the SDP method is used for passive correction for each sub-block. For the complex pole fraction

[0081]

[0082] By constraining Ensure that the sub-block is passive, and then integrate all the sub-blocks to obtain a full frequency domain passive model.

[0083] b. Correction effect verification: The eigenvalue detection of Hamilton matrix is used to ensure that the eigenvalues of the real part of the corrected model are non-negative, and the out-of-bound frequency band of passivity is completely eliminated.

[0084] 4) Simulation efficiency improvement module a. Model compression: based on singular value decomposition of the residual matrix, the rank

[0085] (N is the number of ports) of the residual matrix is compressed, and the R k =U∑V T Simplify the model structure and reduce the number of state variables.

[0086] b. Block parallel simulation: the wideband equivalent model is divided into k sub-modules, and the equivalent admittance and historical current are calculated independently for each sub-module, and the calculation amount of a single sub-module is O component =2nN 2 / k+N 2 +2nN / k.

[0087] The present application includes four steps, namely wideband characteristic scanning, frequency fitting and order identification, passive correction, and simulation efficiency improvement, which are linked through data interface to realize the full-process automation of "scanning-fitting-correction-simulation".

[0088] Interaction interface: Set data input area (power grid parameters, frequency range), parameter configuration area (fitting accuracy, correction threshold), result display area (frequency characteristic curve, error analysis, characteristic value distribution) and operation button (start scanning, fitting calculation, correction execution, simulation acceleration), support batch data processing and log recording.

[0089] Thus, the following effects can be achieved:

[0090] (1) Improve the efficiency of wideband characteristic analysis: The wideband characteristic scanning program developed based on ADPSS can process a 25323-node power grid, and the port characteristic calculation of 500 frequency points takes only 35.5 seconds, which is more than 60% more efficient than the traditional method; The order identification method based on Prony analysis takes 0.13 seconds, which is 96% more efficient than the manual trial method (3.16 seconds).

[0091] (2) Ensure the accuracy of the equivalent: After using the vector fitting method and order identification of the application, the wideband error of mixed simulation and full electromagnetic transient simulation is less than 10%, and in the fault simulation of AC / DC hybrid system, the key indicators such as DC off angle and voltage are consistent with the full electromagnetic transient results, with a consistency of more than 95%.

[0092] (3) Enhance the stability of simulation: The block SDP passive correction method ensures full-frequency-domain passivity, and the characteristic values of the corrected model are all non-negative, and there is no divergence phenomenon in time domain simulation; The total time of passive correction for a 25323-node power grid is 1.62 seconds, which is more than 60% more efficient than the traditional whole SDP method (memory overflow).

[0093] (4) Improve the simulation scale: Through compression and block parallel technology, the module can support high-order network simulation of more than 10 ports, and the single time step calculation amount is reduced by 40%, ensuring the feasibility of real-time simulation of large-scale power grids.

[0094] Optionally, the power grid parameters and the frequency range are obtained, the frequency domain characteristics of the core elements are calculated, and the network boundary is contracted by the Gaussian elimination method, and the multi-port admittance matrix is calculated, including:

[0095] Obtain the power grid parameters and the frequency range, respectively establish the corresponding frequency domain characteristic model for different core elements, and calculate the frequency domain characteristics of the core elements;

[0096] Read the power grid topology and parameters, construct the positive sequence / negative sequence / zero sequence admittance matrix, consider the influence of load and DC device, solve the port frequency domain characteristics by LU decomposition, and output the full frequency domain port admittance sampling data.

[0097] Optionally, the multi-port admittance matrix is fitted, the frequency domain characteristics are converted into time domain sequences by inverse fast Fourier transform, a coefficient matrix is constructed according to the time domain sequences, and the matrix rank is estimated based on singular value decomposition, in combination with the number of real poles, to determine the broadband equivalent model, including:

[0098] Based on the relaxation auxiliary function and by QR decomposition, the efficiency of least square solution is optimized, the rational function of the multi-port admittance matrix Y(s) is fitted, and the expression is:

[0099]

[0100] Wherein, Ri is a residual matrix, a i is a pole, D and E are constant and first-order matrices;

[0101] The constant, first-order and real pole contributions in the frequency domain characteristics are removed, the negative frequency conjugate symmetry data are completed, the time domain sequences are obtained through IFFT, the order is identified through singular value threshold, and the final order is the sum of the number of real poles and the rank.

[0102] Optionally, the broadband equivalent model is split into real pole fractions, constant / first-order part and complex pole fraction sub-blocks, each fraction sub-block is corrected for passivity, and the corrected broadband equivalent model is determined, including:

[0103] The complex pole fraction is corrected according to the following formula

[0104]

[0105] By constraining Ensure the passivity of the sub-block, and integrate all the sub-blocks to obtain a full frequency domain passive model.

[0106] By Hamilton matrix eigenvalue detection, it is ensured that the real part eigenvalues of the corrected broadband equivalent model are non-negative, and the out-of-bound frequency band is eliminated.

[0107] Optionally, the corrected broadband equivalent model is compressed and split into multiple sub-modules for parallel simulation, including:

[0108] Based on singular value decomposition of the residual matrix, the residual matrix of rank (N is the number of ports) is compressed, and the residual matrix is expressed as: k R T Simplify the model structure and reduce the number of state variables.

[0109] The corrected broadband equivalent model is split into k sub-modules, each sub-module independently calculates the equivalent admittance and the history current, and the single time step calculation amount is reduced by parallel calculation, and the calculation amount of a single sub-module is

[0110] Ocomponent = 2nN 2 / k+N 2 +2nN / k.

[0111] Thus, the port frequency characteristics of a large-scale multi-port power grid are accurately sampled, the problem that the order needs to be artificially given in the vector fitting method is solved, the wideband equivalent circuit is ensured to be passive in the full frequency domain and the correction efficiency is improved, the compression and block parallel simulation technology is developed, the real-time simulation efficiency of the wideband equivalent circuit is improved, and the integrated analysis module is constructed to realize integrated operation of wideband characteristic scanning, fitting, correction and simulation acceleration.

[0112] According to another aspect of the application, a system 600 for simulating a wideband equivalent model is also provided, comprising:

[0113] The multi-port admittance matrix calculation module 610 is configured to obtain power grid parameters and a frequency range, calculate frequency domain characteristics of core elements, contract network boundaries based on the frequency domain characteristics by using a Gaussian elimination method, and calculate a multi-port admittance matrix.

[0114] The wideband equivalent model determination module 620 is configured to fit the multi-port admittance matrix, convert the frequency domain characteristics into a time domain sequence by using an inverse fast Fourier transform, construct a coefficient matrix according to the time domain sequence, estimate a matrix rank based on singular value decomposition, combine the number of real poles, and determine a wideband equivalent model.

[0115] The wideband equivalent model correction module 630 is configured to split the wideband equivalent model into real pole fractions, constant term / linear term parts, and complex pole fraction sub-blocks, perform passivity correction on each fraction sub-block, and determine a corrected wideband equivalent model.

[0116] The wideband equivalent model simulation module 640 is configured to compress the corrected wideband equivalent model and split the model into multiple sub-modules for parallel simulation.

[0117] Optionally, the multi-port admittance matrix calculation module comprises:

[0118] The frequency domain characteristic calculation sub-module is configured to obtain power grid parameters and a frequency range, establish a corresponding frequency domain characteristic model for different core elements, and calculate frequency domain characteristics of the core elements.

[0119] The output admittance sampling data sub-module is configured to read power grid topology and parameters, construct positive / negative / zero sequence admittance matrices, take into account the influence of loads and DC devices, solve port frequency domain characteristics by using LU decomposition, and output full frequency domain port admittance sampling data.

[0120] Optionally, the wideband equivalent model determination module comprises:

[0121] A fitting admittance matrix submodule is configured to fit a rational function of a multi-port admittance matrix Y(s) based on a relaxation auxiliary function and by QR decomposition to optimize the efficiency of a least square solution, and an expression is as follows:

[0122]

[0123] wherein Ri is a residue matrix, a i is a pole, and D and E are constant and linear term matrices;

[0124] An order identification submodule is configured to remove constant, linear and real pole contributions in a frequency domain characteristic, complete negative frequency conjugate symmetry data, obtain a time domain sequence through IFFT, identify an order through a singular value threshold, and finally the order is a sum of a real pole number and a rank.

[0125] Optionally, a correction broadband equivalent model module comprises:

[0126] A complex pole fraction correction submodule is configured to correct a complex pole fraction according to the following formula

[0127]

[0128] By constraining to ensure that a sub-block is passive, and a full frequency domain passive model is obtained by integrating all sub-blocks.

[0129] A matrix eigenvalue detection submodule is configured to detect eigenvalues of a Hamilton matrix to ensure that eigenvalues of a real part of a corrected broadband equivalent model are non-negative, and a passive out-of-bound frequency band is eliminated.

[0130] Optionally, a broadband equivalent model simulation module comprises:

[0131] A compression residue tree matrix submodule is configured to compress a residue matrix with a rank (N is a port number) based on singular value decomposition of the residue matrix, and to obtain R k = U∑V T Simplify a model structure and reduce a number of state variables.

[0132] The corrected broadband equivalent model is split into k sub-modules, each of which independently calculates equivalent admittance and historical current, and a single time step calculation amount is reduced through parallel calculation, and the calculation amount of a single sub-module is

[0133] O component = 2nN 2 / k + N 2 + 2nN / k.

[0134] The system 600 for simulating a broadband equivalent model of an embodiment of the present application corresponds to the method 100 for simulating a broadband equivalent model of another embodiment of the present application, which will not be described herein again.

[0135] Those skilled in the art will appreciate that embodiments of the present application can be supplied as a method, a system, or a computer program product. Accordingly, the present application can take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present application can take the form of a computer program product on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROMs, optical storage devices, etc.) embodying computer readable program code thereon. Embodiments of the present application can be implemented in various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript, etc.

[0136] The present application is described in terms of flowcharts and / or block diagrams according to embodiments of the present application. It will be understood that each block of the flowchart and / or block diagrams, and combinations of blocks in the flowchart and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general purpose computer, special purpose computer, embedded processing system, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart and / or block diagram block or blocks. Figure 1 The flowchart and / or block diagrams can also be implemented by one or more of the following: Figure 1 Means for performing the function specified by one or more of the flowchart or block diagram blocks.

[0137] These computer program instructions can also be stored in a computer- readable memory that can direct a computer or other programmable data processing apparatus to function in a particular manner, such that the instructions stored in the computer-readable memory produce an article of manufacture including instructions which implement the flowchart and / or block diagram block or blocks. Figure 1 The flowchart and / or block diagrams can also be implemented by one or more of the following: Figure 1 Means for performing the function specified by one or more of the flowchart or block diagram blocks.

[0138] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process such that the instructions which execute on the computer or other programmable apparatus provide steps for implementing the flowchart and / or block diagram block or blocks. Figure 1 The flowchart and / or block diagrams can also be implemented by one or more of the following: Figure 1 Means for performing the function specified by one or more of the flowchart or block diagram blocks.

[0139] While the preferred embodiments of the application have been described, additional variations and modifications can be made to these embodiments by those skilled in the art once they have the benefit of the present disclosure without departing from the spirit and scope of the application. Accordingly, it is intended that such additions and modifications be included within the scope of the application. It is the following claims, including any amendments thereto, which define the scope of the application.

[0140] Obviously, numerous modifications and variations of the present application are possible in light of the above teachings. It is therefore to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A method for simulating a broadband equivalent model, characterized in that, include: Obtain power grid parameters and frequency range, calculate the frequency domain characteristics of core components, and based on the frequency domain characteristics, shrink the network boundary using Gaussian elimination to calculate the multi-port admittance matrix; The multi-port admittance matrix is ​​fitted, and the frequency domain characteristics are converted into a time domain sequence through inverse fast Fourier transform. The coefficient matrix is ​​constructed based on the time domain sequence, and the matrix rank is estimated based on singular value decomposition. Combined with the number of real poles, the broadband equivalent model is determined. The broadband equivalent model is divided into real pole fractions, constant / first-order terms, and complex pole fractions. Passive correction is performed on each fraction sub-block to determine the corrected broadband equivalent model. The corrected broadband equivalent model is compressed and split into multiple sub-modules for parallel simulation.

2. The method according to claim 1, characterized in that, Obtain power grid parameters and frequency range, calculate the frequency domain characteristics of core components, and based on these frequency domain characteristics, shrink the network boundary using Gaussian elimination to calculate the multi-port admittance matrix, including: Obtain power grid parameters and frequency range, establish corresponding frequency domain characteristic models for different core components, and calculate the frequency domain characteristics of the core components; Read the power grid topology and parameters, construct positive / negative / zero sequence admittance matrices, take into account the influence of load and DC equipment, solve the port frequency domain characteristics through LU decomposition, and output full-frequency domain port admittance sampling data.

3. The method according to claim 1, characterized in that, The multi-port admittance matrix is ​​fitted, and the frequency domain characteristics are converted into a time domain sequence using the inverse fast Fourier transform. A coefficient matrix is ​​constructed based on the time domain sequence, and the matrix rank is estimated based on singular value decomposition. Combining the number of real poles, a broadband equivalent model is determined, including: Based on a relaxation auxiliary function and by optimizing the least squares solution efficiency through QR decomposition, a rational function of the multi-port admittance matrix Y(s) is fitted, and the expression is: Where Ri is the residue matrix, a i Let D and E be the poles, and let D and E be the matrices of constant terms and linear terms, respectively. The constant term, first term, and contribution of real poles in the frequency domain characteristics are removed, the negative frequency conjugate symmetric data are completed, the time domain sequence is obtained by IFFT, the order is identified by the singular value threshold, and the final order is the sum of the number of real poles and the rank.

4. The method according to claim 1, characterized in that, The broadband equivalent model is decomposed into real pole fractions, constant / first-order terms, and complex pole fraction sub-blocks. Passive correction is performed on each fraction sub-block to determine the corrected broadband equivalent model, including: Correction of complex pole fractions using the following formula. Through constraints Ensure that the sub-blocks are passive, and integrate all sub-blocks to obtain a full-frequency domain passive model; By detecting the eigenvalues ​​of the Hamiltonian matrix, it is ensured that the real part of the eigenvalues ​​of the corrected broadband equivalent model is non-negative, and passive out-of-bounds frequency bands are eliminated.

5. The method according to claim 1, characterized in that, The corrected broadband equivalent model is compressed and divided into multiple sub-modules for parallel simulation, including: Based on the singular value decomposition of the residue matrix, for the rank The residue matrix (where N is the number of ports) is compressed using R. k =U∑V T Simplify the model structure and reduce the number of state variables; The corrected broadband equivalent model is divided into k sub-modules. Each sub-module independently calculates the equivalent admittance and historical current. Parallel computation is used to reduce the computational cost per time step, with the computational cost of a single sub-module being [missing information]. O component =2nN 2 / k+N 2 +2nN / k。 6. A system for simulating broadband equivalent models, characterized in that, include: The multi-port admittance matrix calculation module is used to obtain power grid parameters and frequency range, calculate the frequency domain characteristics of core components, and calculate the multi-port admittance matrix by shrinking the network boundary through Gaussian elimination based on the frequency domain characteristics. A broadband equivalent model module is defined to fit the multi-port admittance matrix. The frequency domain characteristics are converted into a time domain sequence through inverse fast Fourier transform. A coefficient matrix is ​​constructed based on the time domain sequence and the matrix rank is estimated based on singular value decomposition. The broadband equivalent model is determined by combining the number of real poles. The broadband equivalent model correction module is used to decompose the broadband equivalent model into real pole fractions, constant / first-order terms and complex pole fraction sub-blocks, perform passive correction on each fraction sub-block, and determine the corrected broadband equivalent model. The broadband equivalent model simulation module is used to compress the corrected broadband equivalent model and split it into multiple sub-modules for parallel simulation.

7. The system according to claim 6, characterized in that, The module for calculating the multiport admittance matrix includes: The frequency domain characteristic calculation submodule is used to obtain power grid parameters and frequency range, and to establish corresponding frequency domain characteristic models for different core components to calculate the frequency domain characteristics of the core components. The output admittance sampling data submodule is used to read the power grid topology and parameters, construct positive / negative / zero-sequence admittance matrices, take into account the influence of load and DC equipment, solve the port frequency domain characteristics through LU decomposition, and output full-frequency domain port admittance sampling data.

8. The system according to claim 6, characterized in that, The module for determining the broadband equivalent model includes: The admissibility matrix fitting submodule is used to fit a rational function of the multi-port admittance matrix Y(s) based on a relaxation auxiliary function and by optimizing the least-squares solution efficiency through QR decomposition. The expression is as follows: Where Ri is the residue matrix, a i Let D and E be the poles, and let D and E be the matrices of constant terms and linear terms, respectively. The order identification submodule is used to remove constant terms, first-order terms, and contributions from real poles in the frequency domain characteristics, complete the negative frequency conjugate symmetric data, obtain the time domain sequence through IFFT, identify the order through the singular value threshold, and finally the order is the sum of the number of real poles and the rank.

9. The system according to claim 6, characterized in that, The module for correcting the broadband equivalent model includes: The complex pole fraction correction submodule is used to correct complex pole fractions according to the following formula. Through constraints Ensure that the sub-blocks are passive, and integrate all sub-blocks to obtain a full-frequency domain passive model; The detection matrix eigenvalue submodule is used to ensure that the real part of the eigenvalues ​​of the corrected broadband equivalent model is non-negative by detecting the eigenvalues ​​of the Hamiltonian matrix, thus eliminating passive out-of-bounds frequency bands.

10. The system according to claim 6, characterized in that, The broadband equivalent model simulation module includes: The compressed residue matrix submodule is used to perform rank-wise operations based on the singular value decomposition of the residue matrix. The residue matrix (where N is the number of ports) is compressed using R. k =U∑V T Simplify the model structure and reduce the number of state variables; The corrected broadband equivalent model is divided into k sub-modules. Each sub-module independently calculates the equivalent admittance and historical current. Parallel computation is used to reduce the computational cost per time step, with the computational cost of a single sub-module being [missing information]. O component =2nN 2 / k+N 2 +2nN / k。 11. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method as described in any one of claims 1-5.

12. An electronic device, characterized in that, include: The computer-readable storage medium as described in claim 9; and one or more processors for executing the program in the computer-readable storage medium.

Images